In　vivo　experiments　have　found　that　animals　encode　their　current　location　through　spatial　neurons　(such　as　grid　cells,　place　cells,　border　cells,　head-direction　cells,　and　so　on)　that　are　usually　classified　according　to　their　functions　(or　firing　patterns).　And　in　vitro　experiments　revealed　that　there　are　two　types　of　excitatory　neurons　within　medial　entorhinal　cortex　according　to　their　structural　characteristics:　stellate　cells　and　pyramidal　cells.　It　is　still　unclear　that　how　neuronal　types　with　different　functions　correspond　to　multiple　structural　types.　Previous　researches　have　modeled　spatial　neurons　primarily　according　to　their　functional　type.　In　this　paper,　we　attempt　to　model　spatial　neurons　according　to　their　structural　type　and　propose　computational　models　of　stellate　cells　in　layer　II　of　medial　entorhinal　cortex.　We　use　leaky　integrate-and-fire　(LIF)　neurons　to　model　the　soma　of　stellate　cells　which　are　driven　by　feedforward　excitatory　current　inputs.　To　obtain　this　drive　current,　up-stream　velocity-controlled　oscillators　(VCOs)　are　used,　and　the　output　of　VCOs　is　encoded　using　an　inverse　hexagonal　discrete　Fourier　transform　(IHDFT).　At　the　same　time,　this　current　is　regulated　by　a　local　recurrent　network　consisted　of　stellate　cells　and　interneurons　and　by　backprojection　signals　from　down-stream　hippocampus.　Simulation　results　demonstrate　that　our　model　generates　grid-like　and　border-like　firing　patterns.　Our　model　explains　the　causes　of　stellate　cell　discharge　activity　and　gives　a　new　direction　for　modeling　spatial　neurons.　©　2021　IEEE